AN ACCOUNTING OF THE DUST-OBSCURED STAR FORMATION AND ACCRETION HISTORIES OVER THE LAST ∼11 BILLION YEARS

We report on an accounting of the star-formation-and accretion-driven energetics of 24 mu m-detected sources in the Great Observatories Origins Deep Survey-North field. For sources having infrared (IR; 8-1000 mu m) luminosities greater than or similar to 3 x 10(12) L-circle dot when derived by fitting local spectral energy distributions (SEDs) to 24 mu m photometry alone, we find these IR luminosity estimates to be a factor of similar to 4 times larger than those estimated when the SED fitting includes additional 16 and 70 mu m data (and in some cases mid-IR spectroscopy and 850 mu m data). This discrepancy arises from the fact that high-luminosity sources at z >> 0 appear to have far-to mid-IR ratios, as well as aromatic feature equivalent widths, typical of lower luminosity galaxies in the local universe. Using our improved estimates for IR luminosity and active galactic nucleus (AGN) contributions, we investigate the evolution of the IR luminosity density versus redshift arising from star formation and AGN processes alone. We find that, within the uncertainties, the total star-formation-driven IR luminosity density is constant between 1.15 less than or similar to z less than or similar to 2.35, although our results suggest a slightly larger value at z greater than or similar to 2. AGNs appear to account for less than or similar to 18% of the total IR luminosity density integrated between 0 less than or similar to z less than or similar to 2.35, contributing less than or similar to 25% at each epoch. Luminous infrared galaxies (LIRGs; 10(11) L-circle dot <= L-IR < 10(12) L-circle dot) appear to dominate the star formation rate density along with normal star-forming galaxies (L-IR < 10(11) L-circle dot) between 0.6 less than or similar to z less than or similar to 1.15. Once beyond z greater than or similar to 2, the contribution from ultraluminous infrared galaxies (LIR >= 10(12) L-circle dot) becomes comparable with that of LIRGs. Using our improved IR luminosity estimates, we find existing calibrations for UV extinction corrections based on measurements of the UV spectral slope typically overcorrect UV luminosities by a factor of similar to 2, on average, for our sample of 24 mu m-selected sources; accordingly we have derived a new UV extinction correction more appropriate for our sample.